The proposed K25 career development award will allow the candidate to become an independent and successful researcher in the emerging multidisciplinary biomedical computing area armed with a unique set of knowledge and breadth of experience both in computational and biomedical sciences. The University of Iowa with a well-rooted multidisciplinary culture, has newly established the Center of Excellence in Image-Guided Radiation Therapy, a world-class radiation therapy facility. In this excellent environment, the candidate, having a joint appointment at Departments of Electrical &Computer Engineering and Radiation Oncology, will develop his research career through (i) taking courses in radiation therapy and medical imaging;(ii) presenting research results at the most pertinent medical and computer science conferences;(iii) attending research group meetings and clinical evaluation and planning meetings at the Center. The research project of the candidate is to develop novel algorithms, methods, and software tools that make use of the latest advances in computer science and medical imaging for accurate target definition and motion tracking, thus improving the treatment effectiveness of Intensity-Modulate Radiation Therapy (IMRT). Target delineation and intra-fraction organ motion are two major sources that compromise the treatment effectiveness of IMRT. The computational feasibility is accomplished by formulating the target delineation and motion tracking problems as computing an optimal closed set in a weighted directed graph. The novel features of our method will be designed with a continuing focus on the global optimality of the solution. We hypothesize that advanced graph algorithmic and geometric techniques enable accurate target delineation and precise tracking of internal tumor/organ motion, thus improving IMRT effectiveness. The specific aims of the proposed research are as follows: 1) Develop and validate a method for the optimal delineation of single and multiple interacting surfaces in volumetric image data;surfaces with terrain-like, tubular, and closed shapes as well as those with complex topologies will be included. 2) Develop and validate a method for tracking optimal organ motion over the treatment course using 4-D image data. 3) Develop and validate a method for accurate tissue voxel mapping which enables to transfer a 3-D treatment plan at one motion phase to all other phases to form a 4-D treatment plan. Public Health Relevance: We expect the clinical applications of the proposed approaches will have a measurable improvement on IMRT treatment effectiveness, leading to a better local tumor control and a significant increase of cancer survival rate.